Ruthenium-catalysed hydroxycarbonylation of olefins

Article abstract of DOI:10.1039/d0cy02283g

State-of-the-art catalyst systems for hydroxy- and alkoxycarbonylations of olefins make use of palladium complexes. In this work, we report a complementary ruthenium-catalysed hydroxycarbonylation of olefins applying an inexpensive Ru-precursor (Ru₃(CO)₁₂) and PCy₃as a ligand. Crucial for the success of this transformation is the use of hexafluoroisopropanol (HFIP) as the solvent in the presence of an acid co-catalyst (PTSA). Overall, moderate to good yields are obtained using aliphatic olefins including the industrially relevant substrate di-isobutene. This atom-efficient catalytic transformation provides straightforward access to various carboxylic acids from unfunctionalized olefins.

Full text of DOI:10.1039/d0cy02283g

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
COMMUNICATION
Ruthenium-catalysed hydroxycarbonylation of
olefins†
Cite this: Catal. Sci. Technol., 2021,
11, 2026
Ricarda Dühren,a Peter Kucmierczyk,bc Ralf Jackstell,a
Robert Frankebc and Matthias Beller
*
a
Received 27th November 2020,
Accepted 7th March 2021
DOI: 10.1039/d0cy02283g
rsc.li/catalysis
State-of-the-art
catalyst
systems
for
hydroxy-
and
Furthermore,
they
allow
for
100%
atom-efficient
alkoxycarbonylations of olefins make use of palladium
complexes. In this work, we report a complementary ruthenium-
catalysed hydroxycarbonylation of olefins applying an
transformations.
Today's knowledge in this area is largely based on the
original work of Reppe and co-workers.4 They enabled Ni-
catalysed carbonylation of acetylene to acrylic acid already in
the 1930's. Since then, important catalyst developments using
other metals in the presence of a variety of ligands took place
and reaction conditions have been improved.5,6 In addition,
substantial research focused on the synthetic applications of
transition-metal catalysed carbonylation reactions7 as well as
practical implementations. In this respect, an important
example is the palladium-catalysed lucite alpha process to
produce methyl methacrylate (MMA) by carbonylation of
ethylene to methylpropanoate8 in the first step followed by
subsequent condensation with formaldehyde.9 This
monomer is used to produce polymethacrylates (PMMA).
Apart from esters, carboxylic acids are important basic and
fine chemicals, which are used to produce derivatives like
carboxylate salts, anhydrides, and amides. In addition,
carboxylic acids find numerous applications in daily life for
instance in coatings, as food additives, disinfecting products,
and pharmaceuticals.10 Hence, there is a continuing interest
in exploring more efficient and less costly routes.
inexpensive Ru-precursor (Ru3(CO)12
) and PCy3 as a ligand.
Crucial for the success of this transformation is the use of
hexafluoroisopropanol (HFIP) as the solvent in the presence of an
acid co-catalyst (PTSA). Overall, moderate to good yields are
obtained using aliphatic olefins including the industrially relevant
substrate
di-isobutene.
This
atom-efficient
catalytic
transformation provides straightforward access to various
carboxylic acids from unfunctionalized olefins.
Introduction
The transition-metal catalysed functionalization of olefins to
form carbonyl-containing compounds is a powerful tool for
the production of numerous fine and bulk chemicals in
industry and for basic research in academia.1,2 In fact,
besides polymerizations and oxidations, carbonylation
reactions using CO as a C1 building block represent the most
commercially
important
chemical
transformations.
Specifically, aliphatic olefins are converted into the
corresponding aldehydes on a million-ton-scale (>10 million
tons per year)3 by so-called hydroformylation. In general,
these products are converted into secondary products like
alcohols or carboxylic acids. In this respect, direct syntheses
For example, in the past decade several palladium-
catalysed hydroxycarbonylation reactions have been
disclosed.11–13 Notably, in these reactions the price of the
metal (palladium) constitutes an important factor for the
overall process costs. Hence, using other less costly metals,
especially ruthenium, is particularly attractive (see Fig. 1). In
of
carboxylic
acid
derivatives
via
alkoxy-
or
hydroxycarbonylation are also of interest due to the improved
step-economy and preventing additional work up.
general, the mechanism of
a
ruthenium-catalysed
hydroxycarbonylation should be like the state-of-the-art
palladium-based transformations.14–16 However, to the best
a Leibniz-Institut für Katalyse e.V., Albert-Einstein-Straße 29a, 18059 Rostock,
Germany. E-mail: matthias.beller@catalysis.de
b Evonik Operations GmbH, Germany. E-mail: robert.franke@evonik.com
c Lehrstuhl für Theoretische Chemie, Ruhr-Universiät Bochum, 44780 Bochum,
Germany
of
our
knowledge
only
one
example
of
a
hydroxycarbonylation of olefins to acids using water in the
presence of a Ru catalyst is known: in 1989, a patent by Shell
International claimed the alkoxy- and hydroxycarbonylation
of olefins using CO Ru/acid/additives to form esters and
carboxylic acids.17 Aside ruthenium complexes have been
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
d0cy02283g
2026 | Catal. Sci. Technol., 2021, 11, 2026–2030
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Communication
previous work on palladium-catalysed hydroxycarbonylation
including a design of experiments study,24 a mixture of
1-octene/water/solvent (1 : 6.5 : 6.75) was used in the presence
of an acid co-catalyst (p-toluenesulfonic acid; PTSA) at 120 °C
for 3.5 h. To find a suitable reaction environment, we tested
a broad range of solvents (Tables 1 and S1† for all tested
solvents). Surprisingly, only in fluorinated solvents, e.g.
hexafluoroisopropanol (HFIP) and trifluoro acetic acid (TFA)
the desired product is obtained in moderate to high yield
(38–70%). To exclude the involvement of fluoride ions in this
transformation additionally some of the non-fluorinated
solvents were tested in the presence of additional CsF.
However, no desired products were observed. At this point it
is interesting to note that positive solvent effects using
fluorinated alcohols like HFIP are well-known. Often it is
supposed that the strong H-donor ability is the cause for
enhanced reactivity.25,26
Fig. 1 Prices of ruthenium and palladium for May 2019 until May 2020
according to https://www.gold.de/kurse/palladiumpreis/ and https://
www.gold.de/kurse/rutheniumpreis/ (status October 2020).
successfully applied in related reactions of olefins such as
hydroesterification
using
formates,18–20
the
alkoxycarbonylation with CO2 as CO surrogate,21 and CH-
carbonylations of (hetero)arenes.22
Herein, we report that commercially available ruthenium
salts and carbonyl complexes in the presence of basic
monophosphines catalyse the carbonylation of olefins in
fluorinated solvents to give the corresponding carboxylic
acids in a chemoselective manner. We believe these findings
will stimulate the use of alternative (less expensive) catalyst
metals in carbonylation reactions in general.
Based on these results, we chose to use HFIP as the
solvent for all following reactions. Interestingly, under these
conditions no significant amounts of the corresponding
aldehydes or alcohols are observed. In order to identify the
optimal catalyst more efficiently, potentially active ruthenium
phosphine complexes were generated in situ. To recognize
the best metal precursor, eight different ruthenium salts and
complexes were compared (Table S2†). Among these only
Ru3(CO)12 and Ru(acac)3 led to the desired C9-acids 2 as a
mixture of linear and branched isomers in moderate to high
yields (63–70%). According to GC-measurements the internal
carboxylic acid isomers were detected as a 1 : 1 : 1 mixture. In
contrast to palladium-catalysed alkoxy-carbonylations of
olefins, halide-containing precursors do not favor this
reaction (Table S2,† entries 2–4, 6 and 7). Thus, further
catalytic tests were carried out with 5 mol% Ru3(CO)12 and
different ligands (Ru : L = 1 : 1.1 for monodentate ligands and
1 : 0.55 for bidentate ligands to keep the ratio of Ru : P in
solution constant) (Fig. 2). In the presence of bidentate
phosphines including the industrially important ligand L2 as
well as Xantphos L5, lower yields (49–62%) of 2 were obtained.
Results and discussion
At the start of this project, we studied the influence of
different solvents, metal precursors and ligands on the
ruthenium-catalysed hydroxycarbonylation of 1-octene. This
benchmark reaction was chosen due to the industrial interest
in the carbonylation of long chain aliphatic olefins. In fact,
the resulting linear product nonanoic acid (pelargonic acid),
is used in the preparation of plasticizers and lacquers.
Moreover, the corresponding esters are applied as flavorings
and ammonium nonanoate is a known herbicide.23 It should
be noted that due to the possibility of olefin isomerization
reactions, this model system is also demanding. Inspired by
Table 1 Investigation of influence of different solvents on the hydroxycarbonylation of 1-octene
Entry
Solvent
Additive
Yield (2) [%]
n/iso [%]
1
2
3
4
5
6
7
8
HOAc
HFIP
TFA
Trifluoroethanol
Dioxane
Dioxane
Anisole
Anisole
—
—
—
—
0
70
38
0
0
0
—
19/81
25/75
—
—
—
—
5 mol% CsF
—
5 mol% CsF
0
0
—
—
Reaction conditions: 1 (2 mmol), Ru-precursor (5 mol% Ru), PTSA·H2O : PCy3 (3.75 : 1, 20.6 mol%), PCy3 : Ru (1.1 : 1, 5.5 mol%), H2O : 1-octene
(6.5 : 1), HFIP : 1-octene (6.75 : 1), CO (40 bar), 120 °C, 3.5 h, VR = 2 mL. Yield of 2 was determined by GC analysis using hexadecane as internal
standard. n/iso describes the ratio of the linear nonanoic acid to all other C9 acid isomers.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2026–2030 | 2027

View Article Online
Communication
Catalysis Science & Technology
Fig. 3 Conversion and product yields for the ruthenium-catalysed
hydroxycarbonylation of 1-octene in course of the time. Reaction
conditions: 1-octene (33 mmol), Ru-precursor (5 mol% Ru), PTSA·H2O :
PCy3 (3.75 : 1, 20.6 mol%), PCy3 : Ru (1.1 : 1, 5.5 mol%), H2O : 1-octene
(2 : 1), HFIP : 1-octene (6.75 : 1), CO (40 bar), VR = 30 mL. Yields are
given as the sum of all four isomers.
water (75 vs. 60%). Those results clearly show that the ligand
stabilizes the reaction system. Next, the influence of CO
pressure was investigated in between 15 and 60 bars. To keep
the absolute pressure at a minimum of 40 bars either pure
CO or mixtures of CO and N2 were used. Previous work on
ruthenium-catalysed hydroformylations and hydroamino-
methylations showed the positive effect of hydrogen or in situ
generated hydrogen on the performance of the active
ruthenium catalyst.27,28 Inspired by that work, we also
performed experiments with 40 bar of CO and additional 5
bar of H2 and with 35 bar of CO and 5 bar of H2. However, in
this case no significant effect of H2 on the reaction system
could be observed (64 and 57% yield). Applying the optimal
conditions (1.7 mol% Ru-precursor, 5.5 mol% PTSA·H2O, 5.5
mol% PCy3, olefin : HFIP : water = 1 : 6.75 : 2, 40 bar CO) 2 was
achieved in 75% yield. With these general conditions in
hand, we measured the concentration of substrates and
products over time (Fig. 3).
Finally, we tested the reactivity of other aliphatic
olefins under the optimized reaction conditions. As shown
in Table 2, internal olefins such as 2- and 4-octene
reacted in a comparable manner to 1-octene (Table 2,
entries 2–3). Applying the industrial mixture of di-
isobutene gave regioselectively the desired linear acid,
albeit in lower yield (Table 2, entry 4). Shorter and longer
chain substrates gave the corresponding acids in 56–84%
yield (Table 2, entries 5–7). In case of 1,7-octadiene the
reactivity dropped and only traces of products could be
obtained (Table 2, entry 8). In a similar manner, methyl
methacrylate (MMA) proved to be unreactive (Table 2,
entry 9). In contrast, cyclohexene, as an example for cyclic
olefins, reacted well and provided cyclohexanecarboxylic
acid in 63% yield (Table 2, entry 10). Lastly, aromatic
olefins (styrene, 4-fluorostyrene) were tested but did not
underwent the desired transformation, probably because a
polymerization took place instead.
Fig.
2
Effect of different ligands on the ruthenium-catalysed
(2 mmol),
hydroxycarbonylation of 1-octene. General conditions:
1
Ru3(CO)12 (5 mol% Ru), monodentate ligands L1, L4, L6–L13, L16, L18–
L20 (5.5 mol%) or bidentate ligands L2, L3, L5, L14–L15, L17 (2.75 mol%),
PTSA·H2O : L (3.75 :1), H2O : 1-octene (6.5 : 1), HFIP : 1-octene (6.75: 1), CO
(40 bar), 120 °C, 3.5 h, VR = 2 mL. Yield of 2 was determined by GC
analysis using hexadecane as internal standard. n/iso describes the ratio
of the linear nonanoic acid to all other C9 acid isomers.
Triphenylphosphine
L1
and
the
water-soluble
triphenylphosphine trisulfonate (TPPTS) ligand L13 also gave
the desired product in moderate yields (34–64%). In contrast
to PPh3, 2-pyridyl-diphenyl-phosphine L4 gave a lower yield
(6%). Interestingly, using cyclohexyl-substituted phosphines
such as easily accessible L6 and L16 gave improved yield (63–
70%) of 2. Because of this positive effect, several easily
accessible heterocyclic monodentate phosphorous ligands
containing dicyclohexyl-phosphino groups L7–L12 and L19
were tested. In general, they led to 2 in moderate to good
yields (58–79%). None of the tested ligands gave improved n :
iso ratios compared to the reaction without ligand. Hence,
the terminal and different internal acids were achieved in
similar amounts. Based on the availability and price of PCy3,
this ligand was used in the following experiments to find out
the optimal concentration of ruthenium, ratio of ligand :
metal and water : substrate (see ESI† Table S3).
Lowering the amount of Ru to 1 mol% resulted in a
significant change of the n/iso ratio towards the iso-products.
Anyways, best results were achieved with 5 mol% of Ru. The
ratio of water : substrate could be lowered to 2 : 1 using PCy3
which leads to an improved space–time-yield. In comparison,
the yield without ligand decreased by lowering the amount of
2028 | Catal. Sci. Technol., 2021, 11, 2026–2030
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Communication
Table 2 Ruthenium-catalysed hydroxycarbonylation of various olefins
Entry
Olefin
Linear acid
Yield [%, n/iso]
75 (22/78)
73 (18/82)
63 (17/83)
29
1
2
3
4
5
6
56 (29/71)
67 (20/80)
7
8
9
84 (22/78)
Traces
0
10
11
63
41
11
12
0
0
General conditions: olefin, Ru3(CO)12 (5 mol% Ru), PTSA·H2O : PCy3 (3.75 : 1, 20.6 mol%), PCy3 : Ru (1.1 : 1, 5.5 mol%), H2O : olefin (2 : 1), HFIP :
olefin (6.75 : 1), CO (40 bar), VR = 2 mL. Yields and ratios of isomers were determined via GC analysis using hexadecane as the internal
standard. n/iso describes the ratio of the linear acid to all other acid isomers.
Conclusions
stirring bar. The vial was closed by PTFE septum and
phenolic cap, fixed in an alloy plate, and connected to
atmosphere with a needle. The vial was evacuated and
recharged with argon for three times. HFIP (1.44 mL), H2O
(0.35 mL) and olefin (2 mmol) were injected and under argon
atmosphere, the vial was transferred into the autoclave. After
the autoclave was flushed three times with 10 bar CO at room
temperature, it was charged with 40 bar CO, and heated to
120 °C for 3.5 hours. Afterwards, the autoclave was cooled to
room temperature and the pressure was released carefully.
The yield was determined by GC analysis with hexadecane
(100 μL) as internal standard.
In conclusion, we have developed a ruthenium-catalysed
hydroxycarbonylation
of
olefins.
Crucial
for
this
transformation is the use of hexafluoroisopropanol as the
solvent. Best results are achieved with the convenient
combination of Ru3(CO)12/PCy3 in the presence of PTSA as
co-catalyst. Various aliphatic olefins (terminal, internal, and
cyclic) provided the corresponding carboxylic acids in
moderate to good yields.
Experimental
General considerations
All air- and moisture-sensitive syntheses were performed
under argon atmosphere in heating gun dried glassware
using Schlenk techniques. Chemicals were purchased from
commercial sources and used as received unless the purity
was less than 98%. O2-Free and dry solvents were prepared
by distillation.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
This work has been funded in part by Evonik Industries, and
the State of Mecklenburg-Western Pomerania and the German
Federal State (BMBF). We thank the analytical team of LIKAT
for their support.
Standard catalytic experiment
In a typical experiment, the vial was charged with Ru3(CO)12
(1.7 mol%), L6 (5.5 mol%), PTSA·H2O (20.6 mol%) and a
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2026–2030 | 2029

View Article Online
Communication
Catalysis Science & Technology
Notes and references
14 K. Dong, R. Sang, X. Fang, R. Franke, A. Spannenberg, H.
Neumann, R. Jackstell and M. Beller, Angew. Chem., Int. Ed.,
2017, 56, 5267–5271.
15 I. del Rio, N. Ruiz, C. Claver, L. A. van der Veen and P. W. N. M.
van Leeuwen, J. Mol. Catal. A: Chem., 2000, 161, 39–48.
16 I. del Rio, C. Claver and P. W. N. M. Van Leeuwen, Eur. J.
Inorg. Chem., 2001, 2719–2738.
17 E. Drent and A. J. M. Breed, EP0329252B1, Shell
International Research, 1989.
18 H. Konishi, T. Ueda, T. Muto and K. Manabe, Org. Lett.,
2012, 14, 4722–4725.
19 S. Ko, Y. Na and S. Chang, J. Am. Chem. Soc., 2002, 124, 750–751.
20 H. Konishi, T. Muto, T. Ueda, Y. Yamada, M. Yamaguchi and
K. Manabe, ChemCatChem, 2015, 7, 836–845.
21 L. Wu, Q. Liu, I. Fleischer, R. Jackstell and M. Beller, Nat.
Commun., 2014, 5, 3091.
1 L. Wu, Q. Liu, R. Jackstell and M. Beller, Angew. Chem., Int.
Ed., 2014, 53, 6310–6320.
2 A. Brennführer, H. Neumann and M. Beller, ChemCatChem,
2009, 1, 28–41.
3 R. Franke, D. Selent and A. Börner, Chem. Rev., 2012, 112,
5675–5732.
4 W. Reppe, O. Schlichting, K. Klager and T. Toepel, Justus
Liebigs Ann. Chem., 1948, 560, 1–92.
5 M. Beller, B. Cornils, C. D. Frohning and C. W.
Kohlpaintner, J. Mol. Catal. A: Chem., 1995, 104, 17–85.
6 G. Kiss, Chem. Rev., 2001, 101, 3435–3456.
7 C. Torborg and M. Beller, Adv. Synth. Catal., 2009, 351,
3027–3043.
8 W. Clegg, G. R. Eastham, M. R. J. Elsegood, R. P. Tooze,
X. L. Wang and K. Whiston, Chem. Commun.,
1999, 1877–1878.
22 A. Tlili, J. Schranck, J. Pospech, H. Neumann and M. Beller,
ChemCatChem, 2014, 6, 1562–1566.
9 J. M. Pearson and R. A. Hadden, US20010051745A1, Lucite
International UK Ltd, 2001.
10 Y. Wang, W. Ren and Y. Shi, Org. Biomol. Chem., 2015, 13,
8416–8419.
11 R. Sang, P. Kucmierczyk, R. Dühren, R. Razzaq, K. Dong, J.
Liu, R. Franke, R. Jackstell and M. Beller, Angew. Chem., Int.
Ed., 2019, 58, 14365–14373.
23 R. A. Smiley, US6323156B1, Falcon Lab LLC, 2000.
24 R. Dühren, P. Kucmierczyk, R. Jackstell, R. Franke and M.
Beller, unpublished work.
25 A. Berkessel, J. A. Adrio, D. Hüttenhain and J. M. Neudörfl,
J. Am. Chem. Soc., 2006, 128, 8421–8426.
26 J. Wencel-Delord and F. Colobert, Org. Chem. Front., 2016, 3,
394–400.
12 V. Goldbach, L. Falivene, L. Caporaso, L. Cavallo and S.
Mecking, ACS Catal., 2016, 6, 8229–8238.
27 L. Wu, I. Fleischer, R. Jackstell and M. Beller, J. Am. Chem.
Soc., 2013, 135, 3989–3996.
13 T. M. Konrad, J. T. Durrani, C. J. Cobley and M. L. Clarke,
Chem. Commun., 2013, 49, 3306–3308.
28 I. Fleischer, L. Wu, I. Profir, R. Jackstell, R. Franke and M.
Beller, Chem. – Eur. J., 2013, 19, 10589–10594.
2030 | Catal. Sci. Technol., 2021, 11, 2026–2030
This journal is © The Royal Society of Chemistry 2021